USB direct or via. a reclocker?

True you could run your mR into the IP2 then into the dac they are in theory fulfilling markedly different roles as Slim says.
Worth trying both together and individually see what you think

1 Like

No, no, a thousand times no.:expressionless:
See this link -

The mR is strictly a transport. How can it clean up the USB signal when it is outputting the signal back out thru the same GEN USB port as the USB port on the originating machine?

Your decision, of course.

I am studiously avoiding the whole argument as to whether a modern asynchronous DAC even needs its USB input ‘cleaned up’. I’ll leave that to the more technically articulate.


So I’ve been using a microRendu for more than 3 months so I know a fair bit about what it is and what it does. If it was ‘just’ a transport then why do Sonore recommend ultra low noise power supplies? Why does John Swenson (designer of the rendu and the LPS-1) explain - at length - the importance of a really short USB connection between the microRendu and the DAC? Why does he also go into detail the importance of ultra low jitter oscillators in the rendu if, at the end of it, it has the same old USB gen as the ROCK? I’m sorry but I’m confused by your conjecture.

Avoiding the controversial viewpoint that most of the things you mention are comments about the device by the designer,, let me address the points you make.

Ultra low noise power supplies would, many believe, be desirable on any device. Even an RPi supposedly benefits from ‘better’ power supplies. You wouldn’t say an RPi ‘cleans’ the signal, would you?

I don’t know the article that you are referring to. Links? None the less, I speculate that the same argument about short USB connections could be made about any device, not just the mR. Again, I’m not sure that’s even true, but I have nothing to dispute that.

Because he designed the bloody thing.

The USB GEN is hardware so referencing the ROCK isn’t really germane, unless you mean the hardware that Roon sells. The Roon folks can say what GEN the USB port is on their machine, but I am saying that the port on the mR and the port on any NUC is the same GEN.

If you have read the link I posted to you, then you will see that JR is subtly and diplomatically saying that the mR is not meant to clean up the signal.

But that’s OK, brother. We seem destined to always disagree with each other.

Well, usbridge cleans the usb signal, and mij w4s dac 2v2 is upgraded with a femtoclock. And indeed these two makes the music more lively and ‘here’. Nevertheless I will try the ifi psu to see if there is further improvement, although I’m already satisfied. In any case I’m satisfied enough to not throw any more money at it, besides the ifi psu.

It’s a known fact that the mRendu is an off-the-shelf SBC with a custom but simple USB board with a standard USB controller/hub. It seems there is nothing inherently much better in it than a rPi. One of the reasons that rPi’s and other SBC’s are popular choices for network endpoints is that they are small, low-power, and thus low-noise. Adding a better PSU further improves the situation if you believe that noise leaches into the DAC via the USB cable and thus affects SQ.

1 Like

I’ve been singing the same song, in this forum, to little effect.

As far as these devices go, the heart wants what the heart wants.

How did you figure it out Slim? You need to let others make their own minds up. I did and what I spent didn’t bankrupt me and was easily recouped. For the record I am not saying some of this has no merit. I think manufacturers are getting better and making these devices unnecessary in some but not all cases.

I get your frustration. We all have different opinions. Ideally we should base our opinion on established facts and direct experience - preferably both. I am not wedded to the microrendu. In fact I’ve had mine up for sale for the past month or so but I may be putting it back into service with the Brooklyn - but then maybe not. I need to do some direct listening with the rendu feeding the Brooklyn compared to a direct USB connection from the ROCK to the Brooklyn. That is the only surefire way I know of coming to a conclusion on such matters.

I’m reminded of a recent experiment I did with my new speakers and whether they should be coupled or decoupled from the stands. After extensive listening in various configurations with isolation between speaker and stand, no isolation, etc. the conclusion I came to was that no isolation and directly coupling the speakers to the stand was by far the best in terms of sound quality. Whilst I was feeling quite happy about this I spoke with the stand manufacturer and he explained that the top plate of the stand (that the speaker sits on) is actually an acoustic damped plate that is two steel plates sandwiching a polymer isolation layer!! So does that mean that my speaker is actually decoupled after all? Who knows and, in a way, I can get too swept up in the minutiae of optimal setups. The manufacturers tend to want to give their customers the best experience- certainly they have more time and knowledge than me - but everyone’s ears, rooms and preferences are different so we have to proceed by trial-and-error ultimately. One thing that I have learned is to follow the manufacturers set up advice.

Not sure how this rambling reponse helps other than to say there are no universal right or wrongs - simply personal preferences and beliefs.

1 Like

As far as I know the mR has basically the same USB circuit as the Uptone Regen, only optimised - as input to the USB circuit is known. The Regen cleans the USB signal. @Rik_Carter is right, the mR does this too.

1 Like

I don’t understand sonics or ethernet theory, like some of the people here, but I do understand, from a tinkerer’s viewpoint, computer hardware. I have built many computers over the years. The first using an 8088 motherboard, in late 1986. I had to buy it from Jameco as this was a decade before Amazon gleamed in Bezo’s eye. At that time, it was much more of a struggle to get stuff working. No forums, etc., other than CompuServe. The board I bought even had a few wires soldered on the backside of the circuit board. A sure sign of a kludge to fix a problem with the original board design.

An ARM based SBC is an ARM based SBC and, my stake in the ground is that, there is very little that can be done to make any appreciable difference.

I save my greatest disdain for those $5000 streamers that are nothing but an ARM SBC, a DVD reader, and some supposedly tweaked music daemons from a Linux fork that probably isn’t being paid the proper licensing fees. Oh yeah, and a beautiful case.

By all means let others make their minds up. I mean to articulate an alternative opinion.

If the cost is no big deal then it’s no big deal, although I think the money could be better spent elsewhere and I understand the pride of ownership.:expressionless:

1 Like

@Rik_Carter, thanks for your reply.

I think we are on the same page as far as the need for some of these endpoints. This disagreement comes in what they are meant to do, but I leave that dog to lie.

What speaker stands are you referring to? Based on the size of my listening room, I have been window shopping Harbeth P3s and a REL sub. The stand for the P3s are of an outrageous price and I am looking to discover a viable alternative.

I’m with you on the monoblocks as drivers, but since I want a tube sound, the Prima Luna Prologue Six is, at least currently, on my list.


You miss my point.

  1. Even if the mR does ‘clean up’ the signal, if and to what extent I won’t re-challenge here, it then puts the signal back out the same type of USB port that the signal originated from.
  2. Others, more technically articulate than I, have argued that there is no signal degradation from using ethernet as a transmission medium. So what is being ‘cleaned up’ exactly? If the answer is the output from the USB port on the sending computer, then I refer you to bullet #1.
  3. I have read, on this forum and elsewhere, long winded technical discussions about whether modern asynchronous DACs even need the USB input ‘cleaned’. The would , of course, include any efforts by the Uptone Regen. I accept the opinion that it isn’t necessary.

It’s the price point, you see. If these devices, and there I include all transports (or if you insist USB cleaner uppers) more that a couple of hundred dollars and most fancy schmancy streamers, were reasonably priced it wouldn’t be such a ripoff.

Really, in the end, I don’t care. Now, I make an oath to my essence. This is absolutely the last time, anywhere, that I bring these points up.


1 Like

Sure it is the price point. I’ve paid far too much for my ‘end point’. mR & upgrade to 1.4, USPCB, Mutec MC3+USB. I guess I could come a long way today with a Pi and a DigiOne. I’d keep the Mutec though. It is very good - and expensive…

1 Like

This is why I hope more manufacturers consider network DACs or network endpoints with fiber inputs. All the ones I’ve seen so far are really expensive.

Hopefully anticipating the Amerikan release of the TEAC NT-505.

@xxx Agreed that ethernet is the preferred method for getting data onto a DAC. Alas and has already been mentioned such DACs are scarce. Most mainstream DACs require USB for high sampling rates, DSD, etc. So I think that is the logic behind the microrendu. Keep the audio data in the ethernet realm for as long as possible before converting it to USB. This is a sensible compromise in the absence of better DAC options at the present time.

It’s fun to read these posts. However, it’s not appropriate for you to assume that Sonore products are the same as a rPi. You were not part of the design team and did no participate in any of the endless discussions about what was done with this product and why. I’m the owner of Sonore and I was.

Yes some parts are similar to other products like all cars have similar components. We are using an off the shelf processor and memory board because it gives us the flexibility to concentrate on what matters. Our main board handles more than just the USB. The main board is the input and output for all signals to and from the processor less onboard memory. This includes the operating system, power to all on board system be it on the main board or the processor board, Ethernet, and clocking to various key components. The goal was simplicity, but even then the design is complicated. Beyond that the difference are in the parts themselves. The quality of the onboard linear supplies, power supply filters, and the quality of the onboard oscillator. Then there are decisions we took on purpose to be different such as not using the USB bus for Ethernet because it causes playback issues at high sample rates. There are other things, but we don’t disclose everything we do and why.

BTW I’m not against people using a rPi if they want to.

1 Like

Ok fair enough, maybe I should have phrased it as there seems to be nothing inherently different than a rPi with respect to sound quality output. I think it’s fairly well agreed that the two demons that haunt usb audio are electrical noise and jitter/timing. How does the selection of chips and design choices you made measurably improve those two aspects vs an rPi or optimized SBC like the allo usbridge? Thanks.

1 Like

Doesn’t do anything more than what a $200 bundle at does.